[pjmlp]

Ironically every time someone proudly assert the existence of F-35 C++ coding standard [0], I am not sure if they actually understand the impact.

The software mess from F-35 would it be even worse without the standard, or has the existence of the standard hardly improved the coding practices as usually gets told.

Not that the answer to this philosophical question solves the issues for everyone affected by the F-35 software problems.

[0] - http://www.stroustrup.com/JSF-AV-rules.pdf

[jiggawatts]

One fun thing about “too big” projects like the F-35 is that the project management overheads cause a kind of recursive overhead, like the rocket equation, but applied to technical outcomes instead of orbital velocity. Any change isn’t “just” the change, it now has to got through review boards, subcontractors, liaisons, integration reviews, etc…

The result is that the F-35 computers are being “upgraded” (lol) to the same compute power as a first-gen Apple Watch… starting this year and finishing who-knows-when.

Meanwhile the F-16 which is “not as important” has already been upgraded with the same kind of chips as modern GPUs and has orders of magnitude more performance than the “flying computer” the the F-35 was supposed to be.

Weep for the poor C++ developers forced to shoehorn modern software into a computer that isn’t yet as powerful as a battery-powered consumer device most people have upgraded three times already.

[FpUser]

>"Weep for the poor C++ developers forced to shoehorn modern software into a computer that isn’t yet as powerful as a battery-powered consumer device most people have upgraded three times already."

You might be surprised by what kind of functionality can be squeezed out of "weak" CPU when programmers know how to work on hardware with limited resources.

[jiggawatts]

Having written 4KB assembly demos for BBS sites back in the 80s and 90s, I would not be surprised at all.

What does surprise me is a $110M plane that is being upgraded at the low-low-cost of a mere $300K each to this: https://www.l3harris.com/all-capabilities/high-performance-i...

Yes, that "High-Performance Integrated Core Processor" is pulling 4.5 kW to produce as much computer power as a typical PC in the late 1990s!

[echoangle]

Is it possible that there is a mistake on the page and they mean 450 Watts?

The "1 ATR SHORT" version lists 2 modules and takes 300 Watts, so 450 Watts would line up perfectly for the "1 ATR LONG" which takes 3 modules. 4.5 kW doesn't make a lot of sense here.

[jiggawatts]

The typo-free spec sheet is an extra $150K.